I\'m trying to write an INSERT INTO that does a some DISTINCT/GROUP BY work. The query runs perfectly fine as a select statement, but will not work if it\'s wrapped into an INS
The problem is solved by automatically changing the value of a parameter (optimizer_features_enable). This value determines the optimizer version of the base, with 11 should not give that problem.
Am I thinking wrong, but is not the sql below equal what you want to achieve?
INSERT INTO MasterRecords(BatchRecordRecordID, SourceID, BatchID)
SELECT DISTINCT RecordID, 101, 150
FROM BatchRecords
WHERE BatchID = 150
;
Wonder if it's an order of execution problem... does it work if you use a CTE? CTE must first materialize thus resolve the group by...
Insert INTO MasterRecords (BatchRecordRecordID, SourceID, BatchID)
WITH BR AS (
SELECT RecordID, 101 AS SourceID, 150 AS BatchID
FROM BatchRecords
GROUP BY RecordID, 101,150)
Select RecordID, SourceID, BatchID FROM BR
or... why the group by and where clause in the first place doesn't seem to be doing anything since recordID isn't an aggregate and isn't part of the group by...
Insert into masterRecords (batchrecordRecordID, SourceID, BatchID) SELECT recordID, 101, 150 from batchRecords
I arrived here trying to solve a similar situation so it seems to me that this kind of problem still appears.
In my case, avoiding any optimizer transformation, did the trick.
I applied a NO_QUERY_TRANSFORMATION hint to the "intoed" SELECT statement and the error disappeared.
In the case of this question, I should rewrite it as:
INSERT INTO MasterRecords
(BatchRecordRecordID, SourceID, BatchID)
SELECT /*+NO_QUERY_TRANSFORMATION*/ RecordID, SourceID, BatchID
FROM (
SELECT RecordID, BatchID, 101 AS SourceID
FROM BatchRecords
WHERE BatchID = 150
GROUP BY RecordID, BatchID
) BR